Linux-2.6.12-rc2v2.6.12-rc2

Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!

1 files changed, 3560 insertions, 0 deletions

diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
new file mode 100644
index 00000000000..092d5fb096b
--- /dev/null
+++ b/fs/xfs/xfs_log.c
@@ -0,0 +1,3560 @@
+/*
+ * Copyright (c) 2000-2004 Silicon Graphics, Inc.  All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * Further, this software is distributed without any warranty that it is
+ * free of the rightful claim of any third person regarding infringement
+ * or the like.  Any license provided herein, whether implied or
+ * otherwise, applies only to this software file.  Patent licenses, if
+ * any, provided herein do not apply to combinations of this program with
+ * other software, or any other product whatsoever.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write the Free Software Foundation, Inc., 59
+ * Temple Place - Suite 330, Boston MA 02111-1307, USA.
+ *
+ * Contact information: Silicon Graphics, Inc., 1600 Amphitheatre Pkwy,
+ * Mountain View, CA  94043, or:
+ *
+ * http://www.sgi.com
+ *
+ * For further information regarding this notice, see:
+ *
+ * http://oss.sgi.com/projects/GenInfo/SGIGPLNoticeExplan/
+ */
+
+/*
+ * High level interface routines for log manager
+ */
+
+#include "xfs.h"
+#include "xfs_macros.h"
+#include "xfs_types.h"
+#include "xfs_inum.h"
+#include "xfs_ag.h"
+#include "xfs_sb.h"
+#include "xfs_log.h"
+#include "xfs_trans.h"
+#include "xfs_dir.h"
+#include "xfs_dmapi.h"
+#include "xfs_mount.h"
+#include "xfs_error.h"
+#include "xfs_log_priv.h"
+#include "xfs_buf_item.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_log_recover.h"
+#include "xfs_bit.h"
+#include "xfs_rw.h"
+#include "xfs_trans_priv.h"
+
+
+#define xlog_write_adv_cnt(ptr, len, off, bytes) \
+	{ (ptr) += (bytes); \
+	  (len) -= (bytes); \
+	  (off) += (bytes);}
+
+/* Local miscellaneous function prototypes */
+STATIC int	 xlog_bdstrat_cb(struct xfs_buf *);
+STATIC int	 xlog_commit_record(xfs_mount_t *mp, xlog_ticket_t *ticket,
+				    xlog_in_core_t **, xfs_lsn_t *);
+STATIC xlog_t *  xlog_alloc_log(xfs_mount_t	*mp,
+				xfs_buftarg_t	*log_target,
+				xfs_daddr_t	blk_offset,
+				int		num_bblks);
+STATIC int	 xlog_space_left(xlog_t *log, int cycle, int bytes);
+STATIC int	 xlog_sync(xlog_t *log, xlog_in_core_t *iclog);
+STATIC void	 xlog_unalloc_log(xlog_t *log);
+STATIC int	 xlog_write(xfs_mount_t *mp, xfs_log_iovec_t region[],
+			    int nentries, xfs_log_ticket_t tic,
+			    xfs_lsn_t *start_lsn,
+			    xlog_in_core_t **commit_iclog,
+			    uint flags);
+
+/* local state machine functions */
+STATIC void xlog_state_done_syncing(xlog_in_core_t *iclog, int);
+STATIC void xlog_state_do_callback(xlog_t *log,int aborted, xlog_in_core_t *iclog);
+STATIC int  xlog_state_get_iclog_space(xlog_t		*log,
+				       int		len,
+				       xlog_in_core_t	**iclog,
+				       xlog_ticket_t	*ticket,
+				       int		*continued_write,
+				       int		*logoffsetp);
+STATIC void xlog_state_put_ticket(xlog_t	*log,
+				  xlog_ticket_t *tic);
+STATIC int  xlog_state_release_iclog(xlog_t		*log,
+				     xlog_in_core_t	*iclog);
+STATIC void xlog_state_switch_iclogs(xlog_t		*log,
+				     xlog_in_core_t *iclog,
+				     int		eventual_size);
+STATIC int  xlog_state_sync(xlog_t *log, xfs_lsn_t lsn, uint flags);
+STATIC int  xlog_state_sync_all(xlog_t *log, uint flags);
+STATIC void xlog_state_want_sync(xlog_t	*log, xlog_in_core_t *iclog);
+
+/* local functions to manipulate grant head */
+STATIC int  xlog_grant_log_space(xlog_t		*log,
+				 xlog_ticket_t	*xtic);
+STATIC void xlog_grant_push_ail(xfs_mount_t	*mp,
+				int		need_bytes);
+STATIC void xlog_regrant_reserve_log_space(xlog_t	 *log,
+					   xlog_ticket_t *ticket);
+STATIC int xlog_regrant_write_log_space(xlog_t		*log,
+					 xlog_ticket_t  *ticket);
+STATIC void xlog_ungrant_log_space(xlog_t	 *log,
+				   xlog_ticket_t *ticket);
+
+
+/* local ticket functions */
+STATIC void		xlog_state_ticket_alloc(xlog_t *log);
+STATIC xlog_ticket_t	*xlog_ticket_get(xlog_t *log,
+					 int	unit_bytes,
+					 int	count,
+					 char	clientid,
+					 uint	flags);
+STATIC void		xlog_ticket_put(xlog_t *log, xlog_ticket_t *ticket);
+
+/* local debug functions */
+#if defined(DEBUG) && !defined(XLOG_NOLOG)
+STATIC void	xlog_verify_dest_ptr(xlog_t *log, __psint_t ptr);
+STATIC void	xlog_verify_grant_head(xlog_t *log, int equals);
+STATIC void	xlog_verify_iclog(xlog_t *log, xlog_in_core_t *iclog,
+				  int count, boolean_t syncing);
+STATIC void	xlog_verify_tail_lsn(xlog_t *log, xlog_in_core_t *iclog,
+				     xfs_lsn_t tail_lsn);
+#else
+#define xlog_verify_dest_ptr(a,b)
+#define xlog_verify_grant_head(a,b)
+#define xlog_verify_iclog(a,b,c,d)
+#define xlog_verify_tail_lsn(a,b,c)
+#endif
+
+int		xlog_iclogs_empty(xlog_t *log);
+
+#ifdef DEBUG
+int xlog_do_error = 0;
+int xlog_req_num  = 0;
+int xlog_error_mod = 33;
+#endif
+
+#define XLOG_FORCED_SHUTDOWN(log)	(log->l_flags & XLOG_IO_ERROR)
+
+/*
+ * 0 => disable log manager
+ * 1 => enable log manager
+ * 2 => enable log manager and log debugging
+ */
+#if defined(XLOG_NOLOG) || defined(DEBUG)
+int   xlog_debug = 1;
+xfs_buftarg_t *xlog_target;
+#endif
+
+#if defined(XFS_LOG_TRACE)
+
+void
+xlog_trace_loggrant(xlog_t *log, xlog_ticket_t *tic, xfs_caddr_t string)
+{
+	if (! log->l_grant_trace) {
+		log->l_grant_trace = ktrace_alloc(1024, KM_NOSLEEP);
+		if (! log->l_grant_trace)
+			return;
+	}
+
+	ktrace_enter(log->l_grant_trace,
+		     (void *)tic,
+		     (void *)log->l_reserve_headq,
+		     (void *)log->l_write_headq,
+		     (void *)((unsigned long)log->l_grant_reserve_cycle),
+		     (void *)((unsigned long)log->l_grant_reserve_bytes),
+		     (void *)((unsigned long)log->l_grant_write_cycle),
+		     (void *)((unsigned long)log->l_grant_write_bytes),
+		     (void *)((unsigned long)log->l_curr_cycle),
+		     (void *)((unsigned long)log->l_curr_block),
+		     (void *)((unsigned long)CYCLE_LSN(log->l_tail_lsn)),
+		     (void *)((unsigned long)BLOCK_LSN(log->l_tail_lsn)),
+		     (void *)string,
+		     (void *)((unsigned long)13),
+		     (void *)((unsigned long)14),
+		     (void *)((unsigned long)15),
+		     (void *)((unsigned long)16));
+}
+
+void
+xlog_trace_iclog(xlog_in_core_t *iclog, uint state)
+{
+	pid_t pid;
+
+	pid = current_pid();
+
+	if (!iclog->ic_trace)
+		iclog->ic_trace = ktrace_alloc(256, KM_SLEEP);
+	ktrace_enter(iclog->ic_trace,
+		     (void *)((unsigned long)state),
+		     (void *)((unsigned long)pid),
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0,
+		     (void *)0);
+}
+
+#else
+#define	xlog_trace_loggrant(log,tic,string)
+#define	xlog_trace_iclog(iclog,state)
+#endif /* XFS_LOG_TRACE */
+
+/*
+ * NOTES:
+ *
+ *	1. currblock field gets updated at startup and after in-core logs
+ *		marked as with WANT_SYNC.
+ */
+
+/*
+ * This routine is called when a user of a log manager ticket is done with
+ * the reservation.  If the ticket was ever used, then a commit record for
+ * the associated transaction is written out as a log operation header with
+ * no data.  The flag XLOG_TIC_INITED is set when the first write occurs with
+ * a given ticket.  If the ticket was one with a permanent reservation, then
+ * a few operations are done differently.  Permanent reservation tickets by
+ * default don't release the reservation.  They just commit the current
+ * transaction with the belief that the reservation is still needed.  A flag
+ * must be passed in before permanent reservations are actually released.
+ * When these type of tickets are not released, they need to be set into
+ * the inited state again.  By doing this, a start record will be written
+ * out when the next write occurs.
+ */
+xfs_lsn_t
+xfs_log_done(xfs_mount_t	*mp,
+	     xfs_log_ticket_t	xtic,
+	     void		**iclog,
+	     uint		flags)
+{
+	xlog_t		*log    = mp->m_log;
+	xlog_ticket_t	*ticket = (xfs_log_ticket_t) xtic;
+	xfs_lsn_t	lsn	= 0;
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	if (!xlog_debug && xlog_target == log->l_targ)
+		return 0;
+#endif
+
+	if (XLOG_FORCED_SHUTDOWN(log) ||
+	    /*
+	     * If nothing was ever written, don't write out commit record.
+	     * If we get an error, just continue and give back the log ticket.
+	     */
+	    (((ticket->t_flags & XLOG_TIC_INITED) == 0) &&
+	     (xlog_commit_record(mp, ticket,
+				 (xlog_in_core_t **)iclog, &lsn)))) {
+		lsn = (xfs_lsn_t) -1;
+		if (ticket->t_flags & XLOG_TIC_PERM_RESERV) {
+			flags |= XFS_LOG_REL_PERM_RESERV;
+		}
+	}
+
+
+	if ((ticket->t_flags & XLOG_TIC_PERM_RESERV) == 0 ||
+	    (flags & XFS_LOG_REL_PERM_RESERV)) {
+		/*
+		 * Release ticket if not permanent reservation or a specifc
+		 * request has been made to release a permanent reservation.
+		 */
+		xlog_ungrant_log_space(log, ticket);
+		xlog_state_put_ticket(log, ticket);
+	} else {
+		xlog_regrant_reserve_log_space(log, ticket);
+	}
+
+	/* If this ticket was a permanent reservation and we aren't
+	 * trying to release it, reset the inited flags; so next time
+	 * we write, a start record will be written out.
+	 */
+	if ((ticket->t_flags & XLOG_TIC_PERM_RESERV) &&
+	    (flags & XFS_LOG_REL_PERM_RESERV) == 0)
+		ticket->t_flags |= XLOG_TIC_INITED;
+
+	return lsn;
+}	/* xfs_log_done */
+
+
+/*
+ * Force the in-core log to disk.  If flags == XFS_LOG_SYNC,
+ *	the force is done synchronously.
+ *
+ * Asynchronous forces are implemented by setting the WANT_SYNC
+ * bit in the appropriate in-core log and then returning.
+ *
+ * Synchronous forces are implemented with a semaphore.  All callers
+ * to force a given lsn to disk will wait on a semaphore attached to the
+ * specific in-core log.  When given in-core log finally completes its
+ * write to disk, that thread will wake up all threads waiting on the
+ * semaphore.
+ */
+int
+xfs_log_force(xfs_mount_t *mp,
+	      xfs_lsn_t	  lsn,
+	      uint	  flags)
+{
+	int	rval;
+	xlog_t *log = mp->m_log;
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	if (!xlog_debug && xlog_target == log->l_targ)
+		return 0;
+#endif
+
+	ASSERT(flags & XFS_LOG_FORCE);
+
+	XFS_STATS_INC(xs_log_force);
+
+	if ((log->l_flags & XLOG_IO_ERROR) == 0) {
+		if (lsn == 0)
+			rval = xlog_state_sync_all(log, flags);
+		else
+			rval = xlog_state_sync(log, lsn, flags);
+	} else {
+		rval = XFS_ERROR(EIO);
+	}
+
+	return rval;
+
+}	/* xfs_log_force */
+
+/*
+ * Attaches a new iclog I/O completion callback routine during
+ * transaction commit.  If the log is in error state, a non-zero
+ * return code is handed back and the caller is responsible for
+ * executing the callback at an appropriate time.
+ */
+int
+xfs_log_notify(xfs_mount_t	  *mp,		/* mount of partition */
+	       void		  *iclog_hndl,	/* iclog to hang callback off */
+	       xfs_log_callback_t *cb)
+{
+	xlog_t *log = mp->m_log;
+	xlog_in_core_t	  *iclog = (xlog_in_core_t *)iclog_hndl;
+	int	abortflg, spl;
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	if (!xlog_debug && xlog_target == log->l_targ)
+		return 0;
+#endif
+	cb->cb_next = NULL;
+	spl = LOG_LOCK(log);
+	abortflg = (iclog->ic_state & XLOG_STATE_IOERROR);
+	if (!abortflg) {
+		ASSERT_ALWAYS((iclog->ic_state == XLOG_STATE_ACTIVE) ||
+			      (iclog->ic_state == XLOG_STATE_WANT_SYNC));
+		cb->cb_next = NULL;
+		*(iclog->ic_callback_tail) = cb;
+		iclog->ic_callback_tail = &(cb->cb_next);
+	}
+	LOG_UNLOCK(log, spl);
+	return abortflg;
+}	/* xfs_log_notify */
+
+int
+xfs_log_release_iclog(xfs_mount_t *mp,
+		      void	  *iclog_hndl)
+{
+	xlog_t *log = mp->m_log;
+	xlog_in_core_t	  *iclog = (xlog_in_core_t *)iclog_hndl;
+
+	if (xlog_state_release_iclog(log, iclog)) {
+		xfs_force_shutdown(mp, XFS_LOG_IO_ERROR);
+		return(EIO);
+	}
+
+	return 0;
+}
+
+/*
+ *  1. Reserve an amount of on-disk log space and return a ticket corresponding
+ *	to the reservation.
+ *  2. Potentially, push buffers at tail of log to disk.
+ *
+ * Each reservation is going to reserve extra space for a log record header.
+ * When writes happen to the on-disk log, we don't subtract the length of the
+ * log record header from any reservation.  By wasting space in each
+ * reservation, we prevent over allocation problems.
+ */
+int
+xfs_log_reserve(xfs_mount_t	 *mp,
+		int		 unit_bytes,
+		int		 cnt,
+		xfs_log_ticket_t *ticket,
+		__uint8_t	 client,
+		uint		 flags)
+{
+	xlog_t		*log = mp->m_log;
+	xlog_ticket_t	*internal_ticket;
+	int		retval;
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	if (!xlog_debug && xlog_target == log->l_targ)
+		return 0;
+#endif
+	retval = 0;
+	ASSERT(client == XFS_TRANSACTION || client == XFS_LOG);
+	ASSERT((flags & XFS_LOG_NOSLEEP) == 0);
+
+	if (XLOG_FORCED_SHUTDOWN(log))
+		return XFS_ERROR(EIO);
+
+	XFS_STATS_INC(xs_try_logspace);
+
+	if (*ticket != NULL) {
+		ASSERT(flags & XFS_LOG_PERM_RESERV);
+		internal_ticket = (xlog_ticket_t *)*ticket;
+		xlog_grant_push_ail(mp, internal_ticket->t_unit_res);
+		retval = xlog_regrant_write_log_space(log, internal_ticket);
+	} else {
+		/* may sleep if need to allocate more tickets */
+		internal_ticket = xlog_ticket_get(log, unit_bytes, cnt,
+						  client, flags);
+		*ticket = internal_ticket;
+		xlog_grant_push_ail(mp,
+				    (internal_ticket->t_unit_res *
+				     internal_ticket->t_cnt));
+		retval = xlog_grant_log_space(log, internal_ticket);
+	}
+
+	return retval;
+}	/* xfs_log_reserve */
+
+
+/*
+ * Mount a log filesystem
+ *
+ * mp		- ubiquitous xfs mount point structure
+ * log_target	- buftarg of on-disk log device
+ * blk_offset	- Start block # where block size is 512 bytes (BBSIZE)
+ * num_bblocks	- Number of BBSIZE blocks in on-disk log
+ *
+ * Return error or zero.
+ */
+int
+xfs_log_mount(xfs_mount_t	*mp,
+	      xfs_buftarg_t	*log_target,
+	      xfs_daddr_t	blk_offset,
+	      int		num_bblks)
+{
+	if (!(mp->m_flags & XFS_MOUNT_NORECOVERY))
+		cmn_err(CE_NOTE, "XFS mounting filesystem %s", mp->m_fsname);
+	else {
+		cmn_err(CE_NOTE,
+			"!Mounting filesystem \"%s\" in no-recovery mode.  Filesystem will be inconsistent.",
+			mp->m_fsname);
+		ASSERT(XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY);
+	}
+
+	mp->m_log = xlog_alloc_log(mp, log_target, blk_offset, num_bblks);
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	if (!xlog_debug) {
+		cmn_err(CE_NOTE, "log dev: %s", XFS_BUFTARG_NAME(log_target));
+		return 0;
+	}
+#endif
+	/*
+	 * skip log recovery on a norecovery mount.  pretend it all
+	 * just worked.
+	 */
+	if (!(mp->m_flags & XFS_MOUNT_NORECOVERY)) {
+		int	error;
+		vfs_t	*vfsp = XFS_MTOVFS(mp);
+		int	readonly = (vfsp->vfs_flag & VFS_RDONLY);
+
+		if (readonly)
+			vfsp->vfs_flag &= ~VFS_RDONLY;
+
+		error = xlog_recover(mp->m_log, readonly);
+
+		if (readonly)
+			vfsp->vfs_flag |= VFS_RDONLY;
+		if (error) {
+			cmn_err(CE_WARN, "XFS: log mount/recovery failed: error %d", error);
+			xlog_unalloc_log(mp->m_log);
+			return error;
+		}
+	}
+
+	/* Normal transactions can now occur */
+	mp->m_log->l_flags &= ~XLOG_ACTIVE_RECOVERY;
+
+	/* End mounting message in xfs_log_mount_finish */
+	return 0;
+}	/* xfs_log_mount */
+
+/*
+ * Finish the recovery of the file system.  This is separate from
+ * the xfs_log_mount() call, because it depends on the code in
+ * xfs_mountfs() to read in the root and real-time bitmap inodes
+ * between calling xfs_log_mount() and here.
+ *
+ * mp		- ubiquitous xfs mount point structure
+ */
+int
+xfs_log_mount_finish(xfs_mount_t *mp, int mfsi_flags)
+{
+	int	error;
+
+	if (!(mp->m_flags & XFS_MOUNT_NORECOVERY))
+		error = xlog_recover_finish(mp->m_log, mfsi_flags);
+	else {
+		error = 0;
+		ASSERT(XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY);
+	}
+
+	return error;
+}
+
+/*
+ * Unmount processing for the log.
+ */
+int
+xfs_log_unmount(xfs_mount_t *mp)
+{
+	int		error;
+
+	error = xfs_log_unmount_write(mp);
+	xfs_log_unmount_dealloc(mp);
+	return (error);
+}
+
+/*
+ * Final log writes as part of unmount.
+ *
+ * Mark the filesystem clean as unmount happens.  Note that during relocation
+ * this routine needs to be executed as part of source-bag while the
+ * deallocation must not be done until source-end.
+ */
+
+/*
+ * Unmount record used to have a string "Unmount filesystem--" in the
+ * data section where the "Un" was really a magic number (XLOG_UNMOUNT_TYPE).
+ * We just write the magic number now since that particular field isn't
+ * currently architecture converted and "nUmount" is a bit foo.
+ * As far as I know, there weren't any dependencies on the old behaviour.
+ */
+
+int
+xfs_log_unmount_write(xfs_mount_t *mp)
+{
+	xlog_t		 *log = mp->m_log;
+	xlog_in_core_t	 *iclog;
+#ifdef DEBUG
+	xlog_in_core_t	 *first_iclog;
+#endif
+	xfs_log_iovec_t  reg[1];
+	xfs_log_ticket_t tic = NULL;
+	xfs_lsn_t	 lsn;
+	int		 error;
+	SPLDECL(s);
+
+	/* the data section must be 32 bit size aligned */
+	struct {
+	    __uint16_t magic;
+	    __uint16_t pad1;
+	    __uint32_t pad2; /* may as well make it 64 bits */
+	} magic = { XLOG_UNMOUNT_TYPE, 0, 0 };
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	if (!xlog_debug && xlog_target == log->l_targ)
+		return 0;
+#endif
+
+	/*
+	 * Don't write out unmount record on read-only mounts.
+	 * Or, if we are doing a forced umount (typically because of IO errors).
+	 */
+	if (XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY)
+		return 0;
+
+	xfs_log_force(mp, 0, XFS_LOG_FORCE|XFS_LOG_SYNC);
+
+#ifdef DEBUG
+	first_iclog = iclog = log->l_iclog;
+	do {
+		if (!(iclog->ic_state & XLOG_STATE_IOERROR)) {
+			ASSERT(iclog->ic_state & XLOG_STATE_ACTIVE);
+			ASSERT(iclog->ic_offset == 0);
+		}
+		iclog = iclog->ic_next;
+	} while (iclog != first_iclog);
+#endif
+	if (! (XLOG_FORCED_SHUTDOWN(log))) {
+		reg[0].i_addr = (void*)&magic;
+		reg[0].i_len  = sizeof(magic);
+
+		error = xfs_log_reserve(mp, 600, 1, &tic, XFS_LOG, 0);
+		if (!error) {
+			/* remove inited flag */
+			((xlog_ticket_t *)tic)->t_flags = 0;
+			error = xlog_write(mp, reg, 1, tic, &lsn,
+					   NULL, XLOG_UNMOUNT_TRANS);
+			/*
+			 * At this point, we're umounting anyway,
+			 * so there's no point in transitioning log state
+			 * to IOERROR. Just continue...
+			 */
+		}
+
+		if (error) {
+			xfs_fs_cmn_err(CE_ALERT, mp,
+				"xfs_log_unmount: unmount record failed");
+		}
+
+
+		s = LOG_LOCK(log);
+		iclog = log->l_iclog;
+		iclog->ic_refcnt++;
+		LOG_UNLOCK(log, s);
+		xlog_state_want_sync(log, iclog);
+		(void) xlog_state_release_iclog(log, iclog);
+
+		s = LOG_LOCK(log);
+		if (!(iclog->ic_state == XLOG_STATE_ACTIVE ||
+		      iclog->ic_state == XLOG_STATE_DIRTY)) {
+			if (!XLOG_FORCED_SHUTDOWN(log)) {
+				sv_wait(&iclog->ic_forcesema, PMEM,
+					&log->l_icloglock, s);
+			} else {
+				LOG_UNLOCK(log, s);
+			}
+		} else {
+			LOG_UNLOCK(log, s);
+		}
+		if (tic)
+			xlog_state_put_ticket(log, tic);
+	} else {
+		/*
+		 * We're already in forced_shutdown mode, couldn't
+		 * even attempt to write out the unmount transaction.
+		 *
+		 * Go through the motions of sync'ing and releasing
+		 * the iclog, even though no I/O will actually happen,
+		 * we need to wait for other log I/O's that may already
+		 * be in progress.  Do this as a separate section of
+		 * code so we'll know if we ever get stuck here that
+		 * we're in this odd situation of trying to unmount
+		 * a file system that went into forced_shutdown as
+		 * the result of an unmount..
+		 */
+		s = LOG_LOCK(log);
+		iclog = log->l_iclog;
+		iclog->ic_refcnt++;
+		LOG_UNLOCK(log, s);
+
+		xlog_state_want_sync(log, iclog);
+		(void) xlog_state_release_iclog(log, iclog);
+
+		s = LOG_LOCK(log);
+
+		if ( ! (   iclog->ic_state == XLOG_STATE_ACTIVE
+			|| iclog->ic_state == XLOG_STATE_DIRTY
+			|| iclog->ic_state == XLOG_STATE_IOERROR) ) {
+
+				sv_wait(&iclog->ic_forcesema, PMEM,
+					&log->l_icloglock, s);
+		} else {
+			LOG_UNLOCK(log, s);
+		}
+	}
+
+	return 0;
+}	/* xfs_log_unmount_write */
+
+/*
+ * Deallocate log structures for unmount/relocation.
+ */
+void
+xfs_log_unmount_dealloc(xfs_mount_t *mp)
+{
+	xlog_unalloc_log(mp->m_log);
+}
+
+/*
+ * Write region vectors to log.  The write happens using the space reservation
+ * of the ticket (tic).  It is not a requirement that all writes for a given
+ * transaction occur with one call to xfs_log_write().
+ */
+int
+xfs_log_write(xfs_mount_t *	mp,
+	      xfs_log_iovec_t	reg[],
+	      int		nentries,
+	      xfs_log_ticket_t	tic,
+	      xfs_lsn_t		*start_lsn)
+{
+	int	error;
+	xlog_t *log = mp->m_log;
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	if (!xlog_debug && xlog_target == log->l_targ) {
+		*start_lsn = 0;
+		return 0;
+	}
+#endif
+	if (XLOG_FORCED_SHUTDOWN(log))
+		return XFS_ERROR(EIO);
+
+	if ((error = xlog_write(mp, reg, nentries, tic, start_lsn, NULL, 0))) {
+		xfs_force_shutdown(mp, XFS_LOG_IO_ERROR);
+	}
+	return (error);
+}	/* xfs_log_write */
+
+
+void
+xfs_log_move_tail(xfs_mount_t	*mp,
+		  xfs_lsn_t	tail_lsn)
+{
+	xlog_ticket_t	*tic;
+	xlog_t		*log = mp->m_log;
+	int		need_bytes, free_bytes, cycle, bytes;
+	SPLDECL(s);
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	if (!xlog_debug && xlog_target == log->l_targ)
+		return;
+#endif
+	/* XXXsup tmp */
+	if (XLOG_FORCED_SHUTDOWN(log))
+		return;
+	ASSERT(!XFS_FORCED_SHUTDOWN(mp));
+
+	if (tail_lsn == 0) {
+		/* needed since sync_lsn is 64 bits */
+		s = LOG_LOCK(log);
+		tail_lsn = log->l_last_sync_lsn;
+		LOG_UNLOCK(log, s);
+	}
+
+	s = GRANT_LOCK(log);
+
+	/* Also an invalid lsn.  1 implies that we aren't passing in a valid
+	 * tail_lsn.
+	 */
+	if (tail_lsn != 1) {
+		log->l_tail_lsn = tail_lsn;
+	}
+
+	if ((tic = log->l_write_headq)) {
+#ifdef DEBUG
+		if (log->l_flags & XLOG_ACTIVE_RECOVERY)
+			panic("Recovery problem");
+#endif
+		cycle = log->l_grant_write_cycle;
+		bytes = log->l_grant_write_bytes;
+		free_bytes = xlog_space_left(log, cycle, bytes);
+		do {
+			ASSERT(tic->t_flags & XLOG_TIC_PERM_RESERV);
+
+			if (free_bytes < tic->t_unit_res && tail_lsn != 1)
+				break;
+			tail_lsn = 0;
+			free_bytes -= tic->t_unit_res;
+			sv_signal(&tic->t_sema);
+			tic = tic->t_next;
+		} while (tic != log->l_write_headq);
+	}
+	if ((tic = log->l_reserve_headq)) {
+#ifdef DEBUG
+		if (log->l_flags & XLOG_ACTIVE_RECOVERY)
+			panic("Recovery problem");
+#endif
+		cycle = log->l_grant_reserve_cycle;
+		bytes = log->l_grant_reserve_bytes;
+		free_bytes = xlog_space_left(log, cycle, bytes);
+		do {
+			if (tic->t_flags & XLOG_TIC_PERM_RESERV)
+				need_bytes = tic->t_unit_res*tic->t_cnt;
+			else
+				need_bytes = tic->t_unit_res;
+			if (free_bytes < need_bytes && tail_lsn != 1)
+				break;
+			tail_lsn = 0;
+			free_bytes -= need_bytes;
+			sv_signal(&tic->t_sema);
+			tic = tic->t_next;
+		} while (tic != log->l_reserve_headq);
+	}
+	GRANT_UNLOCK(log, s);
+}	/* xfs_log_move_tail */
+
+/*
+ * Determine if we have a transaction that has gone to disk
+ * that needs to be covered. Log activity needs to be idle (no AIL and
+ * nothing in the iclogs). And, we need to be in the right state indicating
+ * something has gone out.
+ */
+int
+xfs_log_need_covered(xfs_mount_t *mp)
+{
+	SPLDECL(s);
+	int		needed = 0, gen;
+	xlog_t		*log = mp->m_log;
+	vfs_t		*vfsp = XFS_MTOVFS(mp);
+
+	if (fs_frozen(vfsp) || XFS_FORCED_SHUTDOWN(mp) ||
+	    (vfsp->vfs_flag & VFS_RDONLY))
+		return 0;
+
+	s = LOG_LOCK(log);
+	if (((log->l_covered_state == XLOG_STATE_COVER_NEED) ||
+		(log->l_covered_state == XLOG_STATE_COVER_NEED2))
+			&& !xfs_trans_first_ail(mp, &gen)
+			&& xlog_iclogs_empty(log)) {
+		if (log->l_covered_state == XLOG_STATE_COVER_NEED)
+			log->l_covered_state = XLOG_STATE_COVER_DONE;
+		else {
+			ASSERT(log->l_covered_state == XLOG_STATE_COVER_NEED2);
+			log->l_covered_state = XLOG_STATE_COVER_DONE2;
+		}
+		needed = 1;
+	}
+	LOG_UNLOCK(log, s);
+	return(needed);
+}
+
+/******************************************************************************
+ *
+ *	local routines
+ *
+ ******************************************************************************
+ */
+
+/* xfs_trans_tail_ail returns 0 when there is nothing in the list.
+ * The log manager must keep track of the last LR which was committed
+ * to disk.  The lsn of this LR will become the new tail_lsn whenever
+ * xfs_trans_tail_ail returns 0.  If we don't do this, we run into
+ * the situation where stuff could be written into the log but nothing
+ * was ever in the AIL when asked.  Eventually, we panic since the
+ * tail hits the head.
+ *
+ * We may be holding the log iclog lock upon entering this routine.
+ */
+xfs_lsn_t
+xlog_assign_tail_lsn(xfs_mount_t *mp)
+{
+	xfs_lsn_t tail_lsn;
+	SPLDECL(s);
+	xlog_t	  *log = mp->m_log;
+
+	tail_lsn = xfs_trans_tail_ail(mp);
+	s = GRANT_LOCK(log);
+	if (tail_lsn != 0) {
+		log->l_tail_lsn = tail_lsn;
+	} else {
+		tail_lsn = log->l_tail_lsn = log->l_last_sync_lsn;
+	}
+	GRANT_UNLOCK(log, s);
+
+	return tail_lsn;
+}	/* xlog_assign_tail_lsn */
+
+
+/*
+ * Return the space in the log between the tail and the head.  The head
+ * is passed in the cycle/bytes formal parms.  In the special case where
+ * the reserve head has wrapped passed the tail, this calculation is no
+ * longer valid.  In this case, just return 0 which means there is no space
+ * in the log.  This works for all places where this function is called
+ * with the reserve head.  Of course, if the write head were to ever
+ * wrap the tail, we should blow up.  Rather than catch this case here,
+ * we depend on other ASSERTions in other parts of the code.   XXXmiken
+ *
+ * This code also handles the case where the reservation head is behind
+ * the tail.  The details of this case are described below, but the end
+ * result is that we return the size of the log as the amount of space left.
+ */
+int
+xlog_space_left(xlog_t *log, int cycle, int bytes)
+{
+	int free_bytes;
+	int tail_bytes;
+	int tail_cycle;
+
+	tail_bytes = BBTOB(BLOCK_LSN(log->l_tail_lsn));
+	tail_cycle = CYCLE_LSN(log->l_tail_lsn);
+	if ((tail_cycle == cycle) && (bytes >= tail_bytes)) {
+		free_bytes = log->l_logsize - (bytes - tail_bytes);
+	} else if ((tail_cycle + 1) < cycle) {
+		return 0;
+	} else if (tail_cycle < cycle) {
+		ASSERT(tail_cycle == (cycle - 1));
+		free_bytes = tail_bytes - bytes;
+	} else {
+		/*
+		 * The reservation head is behind the tail.
+		 * In this case we just want to return the size of the
+		 * log as the amount of space left.
+		 */
+		xfs_fs_cmn_err(CE_ALERT, log->l_mp,
+			"xlog_space_left: head behind tail\n"
+			"  tail_cycle = %d, tail_bytes = %d\n"
+			"  GH   cycle = %d, GH   bytes = %d",
+			tail_cycle, tail_bytes, cycle, bytes);
+		ASSERT(0);
+		free_bytes = log->l_logsize;
+	}
+	return free_bytes;
+}	/* xlog_space_left */
+
+
+/*
+ * Log function which is called when an io completes.
+ *
+ * The log manager needs its own routine, in order to control what
+ * happens with the buffer after the write completes.
+ */
+void
+xlog_iodone(xfs_buf_t *bp)
+{
+	xlog_in_core_t	*iclog;
+	xlog_t		*l;
+	int		aborted;
+
+	iclog = XFS_BUF_FSPRIVATE(bp, xlog_in_core_t *);
+	ASSERT(XFS_BUF_FSPRIVATE2(bp, unsigned long) == (unsigned long) 2);
+	XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1);
+	aborted = 0;
+
+	/*
+	 * Some versions of cpp barf on the recursive definition of
+	 * ic_log -> hic_fields.ic_log and expand ic_log twice when
+	 * it is passed through two macros.  Workaround broken cpp.
+	 */
+	l = iclog->ic_log;
+
+	/*
+	 * Race to shutdown the filesystem if we see an error.
+	 */
+	if (XFS_TEST_ERROR((XFS_BUF_GETERROR(bp)), l->l_mp,
+			XFS_ERRTAG_IODONE_IOERR, XFS_RANDOM_IODONE_IOERR)) {
+		xfs_ioerror_alert("xlog_iodone", l->l_mp, bp, XFS_BUF_ADDR(bp));
+		XFS_BUF_STALE(bp);
+		xfs_force_shutdown(l->l_mp, XFS_LOG_IO_ERROR);
+		/*
+		 * This flag will be propagated to the trans-committed
+		 * callback routines to let them know that the log-commit
+		 * didn't succeed.
+		 */
+		aborted = XFS_LI_ABORTED;
+	} else if (iclog->ic_state & XLOG_STATE_IOERROR) {
+		aborted = XFS_LI_ABORTED;
+	}
+	xlog_state_done_syncing(iclog, aborted);
+	if (!(XFS_BUF_ISASYNC(bp))) {
+		/*
+		 * Corresponding psema() will be done in bwrite().  If we don't
+		 * vsema() here, panic.
+		 */
+		XFS_BUF_V_IODONESEMA(bp);
+	}
+}	/* xlog_iodone */
+
+/*
+ * The bdstrat callback function for log bufs. This gives us a central
+ * place to trap bufs in case we get hit by a log I/O error and need to
+ * shutdown. Actually, in practice, even when we didn't get a log error,
+ * we transition the iclogs to IOERROR state *after* flushing all existing
+ * iclogs to disk. This is because we don't want anymore new transactions to be
+ * started or completed afterwards.
+ */
+STATIC int
+xlog_bdstrat_cb(struct xfs_buf *bp)
+{
+	xlog_in_core_t *iclog;
+
+	iclog = XFS_BUF_FSPRIVATE(bp, xlog_in_core_t *);
+
+	if ((iclog->ic_state & XLOG_STATE_IOERROR) == 0) {
+	  /* note for irix bstrat will need  struct bdevsw passed
+	   * Fix the following macro if the code ever is merged
+	   */
+	    XFS_bdstrat(bp);
+		return 0;
+	}
+
+	xfs_buftrace("XLOG__BDSTRAT IOERROR", bp);
+	XFS_BUF_ERROR(bp, EIO);
+	XFS_BUF_STALE(bp);
+	xfs_biodone(bp);
+	return (XFS_ERROR(EIO));
+
+
+}
+
+/*
+ * Return size of each in-core log record buffer.
+ *
+ * Low memory machines only get 2 16KB buffers.  We don't want to waste
+ * memory here.  However, all other machines get at least 2 32KB buffers.
+ * The number is hard coded because we don't care about the minimum
+ * memory size, just 32MB systems.
+ *
+ * If the filesystem blocksize is too large, we may need to choose a
+ * larger size since the directory code currently logs entire blocks.
+ */
+
+STATIC void
+xlog_get_iclog_buffer_size(xfs_mount_t	*mp,
+			   xlog_t	*log)
+{
+	int size;
+	int xhdrs;
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+	/*
+	 * When logbufs == 0, someone has disabled the log from the FSTAB
+	 * file.  This is not a documented feature.  We need to set xlog_debug
+	 * to zero (this deactivates the log) and set xlog_target to the
+	 * appropriate device.  Only one filesystem may be affected as such
+	 * since this is just a performance hack to test what we might be able
+	 * to get if the log were not present.
+	 */
+	if (mp->m_logbufs == 0) {
+		xlog_debug = 0;
+		xlog_target = log->l_targ;
+		log->l_iclog_bufs = XLOG_MIN_ICLOGS;
+	} else
+#endif
+	{
+		/*
+		 * This is the normal path.  If m_logbufs == -1, then the
+		 * admin has chosen to use the system defaults for logbuffers.
+		 */
+		if (mp->m_logbufs == -1) { 
+			if (xfs_physmem <= btoc(128*1024*1024)) { 
+				log->l_iclog_bufs = XLOG_MIN_ICLOGS; 
+			} else if (xfs_physmem <= btoc(400*1024*1024)) { 
+				log->l_iclog_bufs = XLOG_MED_ICLOGS; 
+			} else {
+				/* 256K with 32K bufs */
+				log->l_iclog_bufs = XLOG_MAX_ICLOGS;
+			}
+		} else
+			log->l_iclog_bufs = mp->m_logbufs;
+
+#if defined(DEBUG) || defined(XLOG_NOLOG)
+		/* We are reactivating a filesystem after it was inactive */
+		if (log->l_targ == xlog_target) {
+			xlog_target = NULL;
+			xlog_debug = 1;
+		}
+#endif
+	}
+
+	/*
+	 * Buffer size passed in from mount system call.
+	 */
+	if (mp->m_logbsize != -1) {
+		size = log->l_iclog_size = mp->m_logbsize;
+		log->l_iclog_size_log = 0;
+		while (size != 1) {
+			log->l_iclog_size_log++;
+			size >>= 1;
+		}
+
+		if (XFS_SB_VERSION_HASLOGV2(&mp->m_sb)) {
+			/* # headers = size / 32K
+			 * one header holds cycles from 32K of data
+			 */
+
+			xhdrs = mp->m_logbsize / XLOG_HEADER_CYCLE_SIZE;
+			if (mp->m_logbsize % XLOG_HEADER_CYCLE_SIZE)
+				xhdrs++;
+			log->l_iclog_hsize = xhdrs << BBSHIFT;
+			log->l_iclog_heads = xhdrs;
+		} else {
+			ASSERT(mp->m_logbsize <= XLOG_BIG_RECORD_BSIZE);
+			log->l_iclog_hsize = BBSIZE;
+			log->l_iclog_heads = 1;
+		}
+		return;
+	}
+
+	/*
+	 * Special case machines that have less than 32MB of memory.
+	 * All machines with more memory use 32KB buffers.
+	 */
+	if (xfs_physmem <= btoc(32*1024*1024)) {
+		/* Don't change; min configuration */
+		log->l_iclog_size = XLOG_RECORD_BSIZE;		/* 16k */
+		log->l_iclog_size_log = XLOG_RECORD_BSHIFT;
+	} else {
+		log->l_iclog_size = XLOG_BIG_RECORD_BSIZE;	/* 32k */
+		log->l_iclog_size_log = XLOG_BIG_RECORD_BSHIFT;
+	}
+
+	/* the default log size is 16k or 32k which is one header sector */
+	log->l_iclog_hsize = BBSIZE;
+	log->l_iclog_heads = 1;
+
+	/*
+	 * For 16KB, we use 3 32KB buffers.  For 32KB block sizes, we use
+	 * 4 32KB buffers.  For 64KB block sizes, we use 8 32KB buffers.
+	 */
+	if (mp->m_sb.sb_blocksize >= 16*1024) {
+		log->l_iclog_size = XLOG_BIG_RECORD_BSIZE;
+		log->l_iclog_size_log = XLOG_BIG_RECORD_BSHIFT;
+		if (mp->m_logbufs == -1) {
+			switch (mp->m_sb.sb_blocksize) {
+			    case 16*1024:			/* 16 KB */
+				log->l_iclog_bufs = 3;
+				break;
+			    case 32*1024:			/* 32 KB */
+				log->l_iclog_bufs = 4;
+				break;
+			    case 64*1024:			/* 64 KB */
+				log->l_iclog_bufs = 8;
+				break;
+